Notes

The goal of this study is to identify how well the Bangladesh national cholera surveillance system captures:

  1. the Bangladeshi population
  2. the Bangladeshi population living in high, medium, and low risk cholera areas

We define cholera surveillance zones as geographies surrounding one of the 22 cholera sentinel hospitals. Surveillance zones are defined as the area enclosed within a circular buffer with a radius of x km surrounding the coordinates of a given sentinel hospital. Subdistrict, district, and tertiary care hospitals may have different radii.

The values we want to estimate include:

  1. percentage of the population living in cholera surveillance zones
  2. number and percentage of infections residing in cholera surveillance zones within the last year
  3. number and percentage of infections residing in cholera surveillance zones that are in high, medium, and low infection risk areas
  4. number and percentage of people living in high, medium, and low infection risk areas that would be captured by cholera surveillance zones

We use the following threshold for defining high, medium, and low risk infection areas:

Identifying the cholera surveillance zone

There are 22 sentinel hospital sites (Table 1).

Table 1: Sentinel hospital IDs and locations.

ID Hospital Division Type Latitude Longitude
1 District Hospital Norshingdi Dhaka district 23.92022 90.71846
2 Adhunik Sadar Hospital Habiganj Sylhet district 24.37346 91.41769
3 District Sadar Hospital Cox’s Bazar Chittagong district 21.44159 91.97683
4 Adhunik Sadar Hospital Naogaon Rajshahi district 24.82262 88.93793
5 General Hospital Patuakhali Barisal tertiary 22.35885 90.32737
6 Adhunik Sadar Hospital Thakurgaon Rangpur district 26.02952 88.47657
7 District Sadar Hospital Satkhira Khulna district 22.69032 89.04679
8 Dhaka Medical College Dhaka Dhaka tertiary 23.72539 90.39701
9 Uttara Adhunik Medical College Hospital Dhaka tertiary 23.87482 90.39675
10 Bangladesh Institute of Tropical and Infectious Diseases Chittagong Chittagong tertiary 22.39274 91.75855
11 General Hospital Tangail Dhaka district 24.26432 89.92484
12 General Hospital Narayanganj Dhaka district 23.62455 90.50121
13 Sadar Hospital Chuadanga Khulna district 23.63517 88.84640
14 General Hospital Meherpur Khulna district 23.77835 88.64240
15 General Hospital Comilla Chittagong district 23.45596 91.18204
16 Upazila Health Complex Chaugachha Jesssore Khulna subdistrict 23.25958 89.02446
17 General Hospital Kusthia Khulna district 23.90120 89.12426
18 Upazila Health Complex Madan Mymensingh subdistrict 24.71669 90.94530
19 Upazila Health Complex Chhatak Sunamganj Sylhet subdistrict 25.03337 91.66869
20 Upazila Health Complex Mathbariya Barisal subdistrict 22.28575 89.95365
21 Upazila Health Complex Bakerganj Barisal subdistrict 22.53881 90.34633
22 Health Complex Shibganj Rajshahi subdistrict 24.68770 88.15844

This is a summary of the populations living within buffers of varying sizes around hospital sentinel surveillance sites (Table 2, Figure 1).

Table 2: Proportion of population living within sentinel hospital buffers.

Buffer (Subdistrict-District-Tertiary in km) Pop Proportion
10-10-30 776309170 4.774018
10-20-30 988122894 6.076596
15-25-35 1215785199 7.476636
20-20-30 1068635761 6.571720

In the absence of better data on health care utilization around the hospital sentinel surveillance sites, we assumed that buffers with radii of 10-10-30km, 10-20-30km, 15-25-35km, 20-20-30km around the hospital could serve as proxies for potential hospital catchment areas (Figure 2). Subdistrict, district, and tertiary care sentinel hospitals could have buffer radii of different sizes. We refer to the joint buffer areas as the “cholera surveillance zone” for the national cholera sentinel surveillance system.

Figure 3: Map of 10-10-30km, 10-20-30km, 15-25-35km, 20-20-30km buffers around (subdistrict-district-tertiary care) sentinel hospital sites (left to right, top row to bottom row, respectively).

Mapping infections with serological data

Using data from a nationally-representative survey across Bangladesh, we developed maps of infection across 5 km x 5km cells in Bangladesh. This map suggests that 4.183715510^{7} infections occurred in Bangladesh over the past year (estimated population size ###) [NEED TO ADD DETAILS ON TIME PERIOD FOR THESE ESTIMATES]. We standardized the infection rate estimates in each cell by the population-weighted mean of the Bangladesh infection rate in that cell. This yielded a relative risk of infection for each grid cell (Figure 4).

Figure 4: Risk of infection relative to population-weighted mean

Determining infection risk categories

We examined the distribution of relative risks (Figure 5) and estimated infections (Figure 6) by grid cell in order to identify thresholds for binning infection risk into categories .

Figure 5: Histogram of the risk of infection relative to population-weighted mean

Figure 6: Histogram of the log total number of infections

We used two sets of fixed thresholds to delineate areas with high, moderate, and low infection risk. The first set of thresholds was based on low, moderate, and high risk (RR) of infection relative to the mean infection rate across Bangladesh adjusted by the population of each cell (i.e. the expected infection rate in the cell). 5 km x 5km grid cells were categorized as low risk if RR < 0.5, moderate risk if RR >= 0.5 and RR < 1.5, and high risk if RR >= 1.5 (Figure 7). The second set of threshold was based on absolute case counts, where grid cells were categorized as low risk if infections < 2000, moderate risk if infections >= 2000 and infections < 5000, and high risk if infections >= 5000 (Figure 8). The number of infection thresholds were based on the 20th and 80th percentiles from the distribution of the relative risk of infection.

## [1] "Distribution of relative risk across by 5 km x 5 km cell"
## 0.1328026 0.6989712 1.026415 1.097483 1.385071 3.336598
## [1] "Distribution of number of infections by 5 km x 5 km cell"
## 8.390758 2260.931 5155.061 6718.67 8748.352 319525.5

Table 3: Population living in each risk category. The population living in high, moderate, and low risk areas according to relative risk and infection thresholds across all of Bangladesh.

Risk Level Population (RR threshold) Population (Infection threshold)
High 22052485 127754236
Moderate 115134174 29636755
Low 25424605 5220273

Figure 7: Cholera risk map as categorized by the risk of infection relative to a population-weighted mean

Figure 8: Cholera risk map as categorized by number of infections

Intersecting infections and the cholera surveillance zone

Figure 9: Infection risk, as categorized by relative risk, within cholera surveillance zones (10-10-30km for subdistrict, district, and tertiary care hospitals)Figure 10: Infection risk, as categorized by relative risk, within cholera surveillance zones (10-20-30km for subdistrict, district, and tertiary care hospitals)Figure 11: Infection risk, as categorized by relative risk, within cholera surveillance zones (15-25-35km for subdistrict, district, and tertiary care hospitals)Figure 12: Infection risk, as categorized by relative risk, within cholera surveillance zones (20-20-30km for subdistrict, district, and tertiary care hospitals)

Figure 13: Infection risk, as categorized by number of infections, within cholera surveillance zones (10-10-30km for subdistrict, district, and tertiary care hospitals)Figure 14: Infection risk, as categorized by number of infections, within cholera surveillance zones (10-20-30km for subdistrict, district, and tertiary care hospitals)Figure 15: Infection risk, as categorized by number of infections, within cholera surveillance zones (15-25-35km for subdistrict, district, and tertiary care hospitals)Figure 16: Infection risk, as categorized by number of infections, within cholera surveillance zones (20-20-30km for subdistrict, district, and tertiary care hospitals)

We overlaid the cholera surveillance zone with the binned infection risk maps to examine the distribution of risks in the surveilled areas. This was done for infection risk as categorized by relative risk (Figure 9 - Figure 12) and by the total number of infections (Figure 13 - Figure 16).

We also identified high risk populations that were not captured by the surveillance zone, as categorized by relative risk (Figure 17 - Figure 18) and by total infections (Figure 19 - Figure 20).

Figure 17: Infection risk, as categorized by relative risk, outside of cholera surveillance zones (10-10-30km for subdistrict, district, and tertiary care hospitals)Figure 21: Infection risk, as categorized by relative risk, outside of cholera surveillance zones (10-20-30km for subdistrict, district, and tertiary care hospitals)Figure 22: Infection risk, as categorized by relative risk, outside of cholera surveillance zones (15-25-35km for subdistrict, district, and tertiary care hospitals)Figure 18: Infection risk, as categorized by relative risk, outside of cholera surveillance zones (20-20-30km for subdistrict, district, and tertiary care hospitals)

Figure 19: Infection risk, as categorized by number of infections, outside of cholera surveillance zones (10-10-30km for subdistrict, district, and tertiary care hospitals)Figure 23: Infection risk, as categorized by number of infections, outside of cholera surveillance zones (10-20-30km for subdistrict, district, and tertiary care hospitals)Figure 24: Infection risk, as categorized by number of infections, outside of cholera surveillance zones (15-25-35km for subdistrict, district, and tertiary care hospitals)Figure 20: Infection risk, as categorized by number of infections, outside of cholera surveillance zones (20-20-30km for subdistrict, district, and tertiary care hospitals)

Populations in the cholera surveillance zone

We examined the number of infections falling within each set of buffer distances. We also calculated the percentage of the buffer population infected, and the percentage of total infections across Bangladesh represented for each set of buffer distances (Table 4).

Table 4: Number and percent infections that may be captured in cholera surveillance zones. The percent infected represents the percentage of infected individuals captured within the cholera surveillance zone out of all infected individuals in Bangladesh.

Buffer size Buffer population Number infected Infections in buffer pop (%) Infections in BGD (%)
10-10-30 776309170 9066520 1.17 21.67
10-20-30 988122894 11621152 1.18 27.78
15-25-35 1215785199 14443776 1.19 34.52
20-20-30 1068635761 12988710 1.22 31.05

We then examined the distribution of relative-risk-based categories (Table 5) and number-of-infection-based categories (Table 6) within the cholera surveillance zone, for each buffer distance.

Table 5: Number and percent infections that may be captured in cholera surveillance zones, categorized by relative risk. The infections in surveillance zone represents the percentage of infected people in high/moderate/low risk grid cells among all infections within the cholera surveillance zone. The population in surveillance zone represents the percentage of people living in high/moderate/low risk grid cells among all people within the cholera surveillance zone. The distribution should across risk categories should sum to 100% for each set of buffer sizes.

Buffer size Risk Category Number Infected % infections in Surveillance Zone Population number % population in Surveillance Zone
10-10-30 High 2830025 31.21 6200421 15.23
10-10-30 Moderate 4823083 53.20 20153739 49.51
10-10-30 Low 1413412 15.59 14350826 35.26
10-20-30 High 3338508 28.73 7267961 14.04
10-20-30 Moderate 6717921 57.81 28468877 55.00
10-20-30 Low 1564724 13.46 16020560 30.95
15-25-35 High 4010512 27.77 8699041 13.87
15-25-35 Moderate 8768322 60.71 36835895 58.74
15-25-35 Low 1664943 11.53 17172964 27.39
20-20-30 High 3796062 29.23 8271142 14.75
20-20-30 Moderate 7627924 58.73 31775217 56.67
20-20-30 Low 1564724 12.05 16020560 28.57

Table 6: Number and percent infections that may be captured in cholera surveillance zones, categorized by number of infections The infections in surveillance zone represents the percentage of infected people in high/moderate/low risk grid cells among all infections within the cholera surveillance zone. The population in surveillance zone represents the percentage of people living in high/moderate/low risk grid cells among all people within the cholera surveillance zone. The distribution should across risk categories should sum to 100% for each set of buffer sizes.

Buffer size Risk Category Number Infected Infections in Surveillance Zone (%) Population in Surveillance Zone (%)
10-10-30 High 8459886.90 93.31 91.54
10-10-30 Moderate 569188.31 6.28 7.94
10-10-30 Low 37444.39 0.41 0.52
10-20-30 High 10395815.46 89.46 86.03
10-20-30 Moderate 1144368.63 9.85 12.51
10-20-30 Low 80968.39 0.70 1.45
15-25-35 High 12640100.42 87.51 83.40
15-25-35 Moderate 1663194.61 11.51 14.77
15-25-35 Low 140481.28 0.97 1.84
20-20-30 High 11523818.98 88.72 85.44
20-20-30 Moderate 1361651.66 10.48 13.09
20-20-30 Low 103238.88 0.79 1.47

Populations at different levels of infection risk in Bangladesh

We sought to describe how well the cholera surveillance zones capture High, Moderate, Low populations across the population of Bangladesh and High, Moderate, Low infections among infections in Bangladesh.

We summarized the percentage of high, moderate, and low infection risk populations in Bangladesh that would be captured by cholera surveillance zones at different buffer sizes when risk was categorized both by relative risk (Table 7) and by the number of infections (Table 8).

Table 7: Number and percent infections that may be captured in Bangladesh, categorized by relative risk. The captured at-risk population represents the percentage of high/moderate/low risk populations captured by the cholera surveillance zone out of all high/moderate/low risk populations in Bangladesh. The captured infections represents the percentage of infections in high/moderate/low risk grid cells among all infections in high/moderate/low risk grid cells across Bangladesh.

Buffer size Risk Category Buffer Pop (Sero) Number Infected Captured At-Risk Pop (%) Captured Infections (%)
10-10-30km High 6200421 2830025 28.12 27.28
10-10-30km Moderate 20153739 4823083 17.50 16.64
10-10-30km Low 14350826 1413412 56.44 57.17
10-20-30km High 7267961 3338508 32.96 32.18
10-20-30km Moderate 28468877 6717921 24.73 23.17
10-20-30km Low 16020560 1564724 63.01 63.29
15-25-35km High 8699041 4010512 39.45 38.66
15-25-35km Moderate 36835895 8768322 31.99 30.25
15-25-35km Low 17172964 1664943 67.54 67.34
20-20-30km High 8271142 3796062 37.51 36.59
20-20-30km Moderate 31775217 7627924 27.60 26.31
20-20-30km Low 16020560 1564724 63.01 63.29

Table 8: Number and percent infections that may be captured in Bangladesh, categorized by number of infections. The captured at-risk population represents the percentage of high/moderate/low risk populations captured by the cholera surveillance zone out of all high/moderate/low risk populations in Bangladesh. The captured infections represents the percentage of infections in high/moderate/low risk grid cells among all infections in high/moderate/low risk grid cells across Bangladesh.

Buffer size Risk Category Buffer Pop (Sero) Number Infected Captured At-Risk Pop (%) Captured Infections (%)
10-10-30km High 37262742.9 8459886.90 29.17 23.98
10-10-30km Moderate 3230533.6 569188.31 10.90 10.07
10-10-30km Low 211709.2 37444.39 4.06 4.14
10-20-30km High 44528498.2 10395815.46 34.85 29.46
10-20-30km Moderate 6476193.0 1144368.63 21.85 20.25
10-20-30km Low 752707.0 80968.39 14.42 8.96
15-25-35km High 52295315.8 12640100.42 40.93 35.83
15-25-35km Moderate 9259598.8 1663194.61 31.24 29.43
15-25-35km Low 1152984.7 140481.28 22.09 15.55
20-20-30km High 47905627.1 11523818.98 37.50 32.66
20-20-30km Moderate 7336363.0 1361651.66 24.75 24.09
20-20-30km Low 824929.5 103238.88 15.80 11.43

(Evaluate according to CDC surveillance guidelines) How representative and sensitive is the system to capturing a proportion of cases?